Due to recent advances in connectivity technology, vehicles in remote areas around the world can connect to the Internet. Data scientists will have the ability to analyze vehicle features and their usage under different driving conditions for the first time, helping to optimize existing designs or offer new features. However, collecting and maintaining big data comes at a considerable cost. Unlike data storage, which tends to be inexpensive, data transmission, data governance, and computational resources are not. Specifically, for connected vehicle data, even when collection of unstructured data such as images is excluded, terabytes of data can potentially still be collected daily. However, due to the various constraints, it is important to apply methods that assist in collecting only what is important for the studies being conducted. In this paper, we argue that Large Language Models play an important role in collecting relevant data for complex studies. As an example, we demonstrate our system that assists various users in choosing a representative sample and provide two case studies to demonstrate the benefits of the system.